MLB's Pitch f/x Data

[rluzinski]

I'm very late to the party but I've started messing with some of the Pitch f/x data that MLB provides through their Gameday service, which can also be accessed here:

 

http://gd2.mlb.com/components/game/mlb/year_2008/

 

I remember a fellow Brewer blogger playing with this stuff as well. Who was that? I want to make sure I'm not wasting my time doing something that someone else already has done (and probably better). I'd also like some questions answered if possible.

 

I was curious where Sheet's fastball was for his first start and how it might have been affected by the rain delay. I am using what they call the starting velocity, which is really the velocity of the ball when it's 50 ft from homeplate. I don't know how that compares to what a radar gun typically picks up. It seems to be in the same ballpark (the ending velocity is significantly lower).

 

Sheets threw 62 fastballs over 6+ innings on Monday. Here are the basics:

 

Ave: 92.4 MPH

Max: 94.6 MPH (third pitch after rain delay)

Min: 87.5 MPH (first pitch of the game)

Median: 92.6 MPH

 

Here's his fastball velocity over the course of the game. The white vertical lines show the start of a new inning. The yellow line is the rain delay:

http://i14.photobucket.com/albums/a345/rluzinski/sheets_small.jpg

I'm sure that everyone knows that it usually takes a little while for Sheets to get his velocity up and this game was no different. Is this pretty typical for a pitcher? Typical for Sheets?

[Chio151]

That curve looks like what we would expect most of the time, but it's fascinating to see it displayed like that. Thank you Russ.

 

I wonder if any other pitchers start out a bit higher and have a consistent downward trend.

[Oldcity]

Interesting stuff. Based solely on observation, it certainly seems like it takes Ben awhile to get going. He has had many starts where he gets knocked around a little in the first inning only to settle down afterwards. I think most pitchers are like that, but it seems like it takes him a bit longer.

[Tbadder]

As a non-stat kinda guy I found it fascinating. I would've guessed he was more inconsistent. I just assumed there was more variation because pitchers intentionally "take something off" ocassionally. Although if they took too much off we'd call it a change-up.

[jazzytrav]

That's one of the biggest reasons I love Gameday. I also like the "pitch break" effect. Not being able to watch the game live, it was nice to actually see the break of Benny's curve, followed by a big red "STRIKE OUT" on the screen.

 

I think this sort of thing would be interesting to do every so often during the course of the season for all of the pitchers. It would be nice to know that all of the games will have equally as reliable numbers, but that might be the only problem right now.

[DrWood]

watching the game on ESPN2, I didn't recall Sheets ever going over 91 mph on his pitches, so I wonder where those numbers come from.

[jazzytrav]

That's why I'd be suspicious as to the reliability of the numbers over the course of a season. As I watched it on Gameday, he was consistently around 93, which is about what rluzinski showed as the median. But then again, that's the pitch speed out of his hand, so it depends where the radar is taking it from.

[RightFieldCoder]

[Russ is] using what they call the starting velocity, which is really the velocity of the ball when it's 50 ft from homeplate. I [Russ] don't know how that compares to what a radar gun typically picks up. It seems to be in the same ballpark (the ending velocity is significantly lower).

I think ESPN2 was using ending velocity.

[rluzinski]

Ending velocity is way less (round 87). Marmol topped out at 94 MPH and Zambrano averaged 91.3 both sound reasonable. A radar gun might be picking up the velocity 25 ft away or somethign like. I think it's the radar guns' accuracy that should be questioned, since the velocity it's recording is from an unknown distance away?

 

Zambrano's velocity curve is the exact opposite Sheets'. It's a big smiley face. His velocity was the highest int he 1st inning, averaging 93.9 MPH. How odd.

[jazzytrav]

Zambrano comes off to me as the kind of guy that's so fired up right out of the gate that his first pitch is mid-90s, then falling off from there.

[rluzinski]

http://i14.photobucket.com/albums/a345/rluzinski/Zambrano.jpg

Ave: 91.3 MPH

Max: 96.2 MPH (7th pitch of the game)

Min: 86.7 MPH

Median: 91.7 MPH

 

Now, pitch type identification isn't perfect (pitch f/x even includes a "type confidnce" column), so maybe a few of those sub 90 MPH fastballs are really something else?

[Ennder]

Some pitchers just change speeds on their fastball. Maybe the sub 90 ones are the ones with so much movement on them.

[jazzytrav]

But why would he change speeds for large portions of time? It would make sense if he changed speeds every couple of pitches just to change it up, but Z was consistently over 90 until around pitch 30 where he suddenly dropped consistently below.

 

You have to admit, there is a distinct difference between 0-30 and 30-60.

[rluzinski]

Yeh, it doesn't seem like he was just changing up speeds but who knows.

 

Do fastballs fatigue a pitcher significantly more than off speed pitches or should I just use overall pitch count? Probably overall pitch count, huh?

[rluzinski]

Sheets' velocity looked a lot better during his second start

Fastball Velocity [MPH]
1st Start / 2nd Start:
AVE: 92.4 / 94.5
MAX: 94.6 / 96.2
MIN: 87.5 / 92.4
MED: 92.6 / 94.6

http://i14.photobucket.com/albums/a345/rluzinski/sheets0406.jpg
If you remove those first two fastballs:

 

http://i14.photobucket.com/albums/a345/rluzinski/sheetsminus20406.jpg

Even if we ignore the first couple of fastballs from Sheets' first start, his fastball was still significantly slower during the early innings.

Looking at pitch locations, Sheets was REALLY getting the benefit of the doubt on outside pitches to LH hitters. While that's not uncommon in the majors:

http://rluzinski.blogspot.com/2007/07/size-and-shape-of-strike-zone-dependant.html

... this was very pronounced. Here are all the called strikes against LH batters:
http://i14.photobucket.com/albums/a345/rluzinski/calledstrikesagainstLH.jpg

With a strikezone like that, it's not surprising that the Giant batters were taking so many defensive swings.

[TooLiveBrew]

I know I've asked this before, but I still don't understand what the difference is between the categories of "Break" & "Pitch f/x". Both are utilized during Gameday broadcasts. I think I know what "break" is -- since I assume it's the common concept, but then what it "pitch f/x", and how is that different from "break"?

[rluzinski]

"Pfx" is the break from 40 ft away and takes out the affects of gravity.

 

"Break" typically starts 50 ft from homeplate and includes gravity.

 

They use different definitions of break though, so I don't think you can really compare one to the other:

 

http://webusers.npl.uiuc.edu/~a-nathan/pob/tracking.htm

[TooLiveBrew]

So "f/x" is just the way the ball breaks due exclusively to its english/spin? Or does a ball not "break" without gravity?

Btw, the link didn't work for me, but that's essentially irrelevant since your explanation is very good. Thanks, rluz.

[rluzinski]

I fixed the link. Fastballs tend to rise (they drop less when gravity is taken into account) and a 12-6 curve will break straight down.

[rluzinski]

I took a look at Sheets' third start, focusing on all those glorious changeups:

 

Ben Sheets' Third Start: Changeup Edition

 

9 changeups. 4 outs, of which were swinging strikes. I wonder if he's going right back to being a two pitch pitcher or will this give him a little confidence with the changeup going forward?

[jazzytrav]

Rluz, I think you need a title. Like "Pitch f/x guru" or something. Those graphs are so cool it's sickening.

 

It's odd to think how well an 86 MPH changeup can work, but it almost seems like that ever-so-slight differential would make even more of a difference if hitters are aware that the pitcher has a low to mid-90s fastball. If there's a 15 MPH difference between fastball and changeup, I would think it would be easier for the hitter to detect changeup. But with the slight change of 6-7 MPH, it seems like it would be nearly impossible to detect. As long as he gets a fastball or two by them first, that changeup could be devastating.

[TuesdaysWithRillo]

I wonder if this new experiment by Sheets will end up like his last experiment with the Cut Fastball. He tried adding that pitch and then he struggled for a few starts and said he wanted to get back to being the pitcher he was: a fastball/curveball pitcher. Don't be surprised if Sheets struggles at some point this year he abandons the changeup and goes back to his bread and butter.

[rluzinski]

Rluz, I think you need a title. Like "Pitch f/x guru" or something. Those graphs are so cool it's sickening.

 

I'm just getting started. There a guys who are basically experts at slicing the data up.

If there's a 15 MPH difference between fastball and changeup, I would think it would be easier for the hitter to detect changeup. But with the slight change of 6-7 MPH, it seems like it would be nearly impossible to detect. As long as he gets a fastball or two by them first, that changeup could be devastating.

 

I think the average is about 7 MPH. As long as you keep the delivery consistent between the fastball and changeup, I always thought that the bigger the difference, the better. I think the batters pickup most of the "velocity data by looking at the pitcher's delivery.

[logan82]

rluzinski wrote:

I think the average is about 7 MPH. As long as you keep the delivery consistent between the fastball and changeup, I always thought that the bigger the difference, the better. I think the batters pickup most of the "velocity data by looking at the pitcher's delivery.

I thought when Wise started to have struggles last year his change and fastball got closer together speed wise. Just as an example.

 

Last year I think we saw Sheets trying all kinds of things since his curve wasn't working for him. He was still pretty effective, but he wasn't at his best by any means.

[jazzytrav]

I wonder if this new experiment by Sheets will end up like his last experiment with the Cut Fastball. He tried adding that pitch and then he struggled for a few starts and said he wanted to get back to being the pitcher he was: a fastball/curveball pitcher. Don't be surprised if Sheets struggles at some point this year he abandons the changeup and goes back to his bread and butter.

I think the key is that he not start using the change to replace his curve in situations where he would normally throw the curve. He's already a good pitcher as a fastball/curveball guy, but tossing in 9 or 10 changeups in there (if he keeps it at that) could make him even better.

 

I think the second graph that Rluz showed (bar graph of pitch speeds) was the coolest because you saw a ton of fastball/curveball, and just a few splashes of changeup.

 

 

I think the average is about 7 MPH. As long as you keep the delivery consistent between the fastball and changeup, I always thought that the bigger the difference, the better.

I'm sorry, you're absolutely right. I think the idea that was really going through my head is that because he has a mid-90s fastball, his changeup should be more effective than someone that throws in the high-80s. That's mostly a gut feeling, but it seems logical that if a guy throws mid-90s heat, the hitter has to be more prepared for the fastball, making the change that much more effective.